3 Methods
3.1 Protein Isolation and Expression
3.1.1 Myosin V expression
Myosin V S1 is expressed (baculovirus system) with the first 792 amino acid residues which includes 1 IQ domain as detailed previously by the Yengo Lab (Gunther et al. 2020). Additionally, expressed myosin V contains the N-terminal tetracysteine motif, C-terminal Myc, and C-terminal FLAG tags (Trivedi et al. 2013, 2015, 2020; Gunther et al. 2019). The S217A mutation was introduced (serine to alanine) using QuikChange site-directed mutagenesis (Stratagene), co-expressed with calmodulin, and purified with FLAG affinity chromatography.
3.1.2 Skeletal muscle myosin II isolation
Fast skeletal muscle myosin II was isolated from chicken pectoralis muscle (Diemand Farm, Wendell, MA). All procedures during isolation were carried out in a cold room or performed on ice. Isolation was performed as previously described by the Debold lab (Woodward et al. 2020; Unger and Debold 2019; Longyear, Walcott, and Debold 2017) by Mike with a protocol similar to those of Margossian and Lowey (1982) with minor modifications. In short, chicken breast muscle is passed through a meat grinder and rinsed with 0.2M EDTA. 2 mL of Buffer A (Extraction buffer consisting of 0.3M KCl, 0.15M KPi, 20mM EDTA, 5mM MgCl2, 3.3mM ATP, and 5mM DTT at pH 6.7) is added per gram of tissue including 5mL of protease inhibitors. This is mixed for 12 minutes with an overhead stirrer. The reaction is stopped with a 4X dilution into water which is then mixed and filtered. After precipitate settles it is centrifuged at 10,800g for 10 minutes at 4C and the resulting pellet is resuspended with Buffer B (Suspension buffer consisting of 1M KCl, 60mM KPi, 20mM EDTA, and 5mM DTT at pH 6.7) and mixed gently before being left to dialyze overnight. Actomyosin is then precipitated and centrifuged at 41,171g for 1 hour at 4C and the resulting supernatant is diluted 10X with water. Clear supernatant is siphoned off and the rest is centrifuged again at 10,800g for 15 minutes at 4C. Supernatant is poured off and precipitate is resuspended with Buffer D (Resupsension buffer consisting of 3M KCl, 50mM KPi, and 5mM DTT at pH 6.7) before being dialized overnight for a second time against Buffer E (Dialysis buffer consisting of 0.6M KCl, 50mM KPi, 1mM NaN3, and 5mM DTT at pH 7.0). After the dialysis, myosin is clarified with an ultracentrifugation at for 2 hours at 4C, concentration determined, snap frozen with liquid nitrogen, and stored at -80C.
3.1.3 Actin isolation and labeling
Acetone powder was prepped from the remainder from the myosin isolation (with the leftovers from the filtered cheesecloth) and actin purification was performed from the resulting acetone powder as described by Pardee and Spudich (1982) with modifications. Briefly, acetone powder was finely ground and mixed with extraction buffer (2mM Tris Base, 0.2mM CaCl2, and 0.005% NaN3 at pH 8.0) and stirred with an effort to minimize creation of bubbles. The resulting solution is spun at 28960g for 20 minutes and supernatant filtered off and kept aside. Additional extraction buffer added to gel-like precipitate and centrifuged a second time with the same specs with the supernatant filtered off and combined with the previous. Actin is polymerized from the resulting supernatant by addition a final polymerization solution (50mM KCl, 2mM MgCl2, and 1mM ATP). Salt is added to slowly to prevent “salt shocking” the proteins before being left to stir overnight. The next day, a high salt wash (increase KCl to 600mM) removes tropomyosin from the f-actin and then then the sedimentation of f-actin performed by centrifuging at 205835g for 60 minutes. The precipitate is transferred to a homogenizer and resuspended with extraction buffer. A 4 day dialysis is performed with extraction buffer additionally containing ATP and DTT to de-polymerize actin. After dialysis the resulting G-actin is clarified with an ultracentrifugation at 200,000g for 60 minutes. After actin is polymerized by adding 10mM Imidazole (pH 7.0) and 1mM MgCl2 and dialyzed against final storage buffer (4mM Imidazole, 25mM KCl, 2mM MgCl2, 1mM NaN3, and 0.01mM ATP at pH 7.0). After calculation of final concentration actin is snap frozen in liquid nitrogen and stored at -80C. After isolation actin can then be labeled with 100% TRITC for use in vitro motility or mixed with a 50:50 TRITC/Biotin solution for use in the three-bead laser trap assay.
3.2 Laser trap assay
The laser trap assay was performed as previously described by the Debold Lab (Woodward et al. 2020; Unger and Debold 2019; Longyear, Walcott, and Debold 2017) with special considerations for the expressed myosin V. Single molecules of myosin were adhered to a nitrocellulose coated microscope slide containing 3µM glass pedestal beads with an additional coverslip glued on top for construction of a “flow-cell.” The final myosin concentration of ~0.8-1µg/mL was added after to introduction of anti-myC antibody (0.8µg/mL, Sigma Inc.) which provided a binding interface for the expressed myosin on the surface. Bovine Serum Albumin (BSA) was used to block the remainder of the surface before the addition of final buffer. The final buffer consisted of an actin buffer (91mM KCl, 1mM EGTA, 4mM MgCl2, and 1mM DTT at pH 7.0) mixed with 100µM ATP, and an oxygen scavenger system (29mM glucose, 1.5mM glucose oxidase, and 80 units catalase) at pH 7.0. For 30mM Pi experiments KCl was reduced in order to maintain the 125mM total ionic strength to match the control 0mM Pi experiments. The concentration of TRITC/Biotin labelled actin filaments and neutravadin/streptavidin coated 1 micron beads (Bangs Lab Inc) was varied at trappers discretion. Bead-actin-bead “dumbbell” setups were constructed using a three axis piezo controlled state (Mad City Labs) with a time shared laser trap between two positions. Experiments were performed at 1.5 Watts laser power and actin filaments pretension to 3-4pN. The resulting system stiffness of the two laser traps and the pretension across the filaments was 0.04pN/nm, determined via the equipartition method (DUPUIS et al. 1997). Bead position was tracked using a four quadrant photodiode with a sampling rate of 5kHz.
3.3 Ultra-fast force clamp
The UFFC data was collected under the same experimental conditions and buffers as described in the section 3.2 for the standard laser trap assay, but full length fast skeletal muscle myosin IIx was adhered to the nitrocellulose coated coverslips instead of the myosin V and anti-body combination. Since UFFC is an extension of the standard three-bead assay implementation was similar in regards to setting up a dumbbell. However, in order to increase the speed that the dumbbell was able to move through solution when the force clamp was engaged smaller 510nm beads (Bangs Lab) were coated with neutravidin (31000, Pierce) for use in the UFFC. Decreasing the radius of the beads reduces the Stoke’s drag coefficient since \(\beta=6 \pi \eta r\), where \(\beta\) represents the drag coefficient, \(\eta\) the viscosity of solution, and r is the radius of the bead. Furthermore, the drag coefficient is inversely proportional to the velocity of the dumbbell setup as given by \(velocity = F_{total}/\beta\). The smaller the drag the faster the dumbbell can move. Consequently, the increased velocity of the dumbbell setup in the un-attached state increases the time resolution and signal-to-noise ratio of the resulting data. Force was pre-determined for each experimental condition and set by calculating the bead’s displacement from the center of the trap and converted to a force since \(F = -kx\), where \(x\) is the bead’s displacement and \(k\) is the stiffness of the trap. The trap stiffness was similar to the standard laser trap experiments at roughly 0.04 pN/nm calculated by analysis of the power spectrum. Positions of the traps were controlled through the AODs and a custom LabView program. Bead position was collected at 200kHz sampling frequency. The total feedback delay was around 8;micro:s, this is the total time it takes for the computers to detect changes in the beads position relative to the center of the trap, communicate the information, and for the AODs to respond to the changes.
Actomyosin interactions in UFFC are identified by applying a threshold the velocity transformation of the raw displacement data that is generated during data collection. The velocity was calculated instantaneously on a point-to-point basis and the results smooth with a Gaussian filter. The velocity transformation results in a double Gaussian distribution with two peaks corresponding to the average velocity of the bound and unbound populations. The bound population velocity is centered around zero because as myosin binds actin and imposes its own stiffness/drag the UFFC will feedback in attempts to apply a consistent force to the dumbbell causing the traps to stop moving. The threshold is set at the point along the joint PDF of the double Gaussian where the probability of crossing the threshold due to noise from the unbound or bound event is equivalent. The threshold was then optimized for each record to decrease the number of false events detected to <1%. If false events exceeded 1%, the SD of the Gaussian filter was increased to further smooth out the data in order to decrease the probability of a baseline noise artifact crossing the threshold. Note that usually smaller SD of the gaussian filters could be applied at great forces due to an increase in force subsequently increasing the signal-to-noise ratio since the baseline velocity is faster to achieve higher forces. Additionally, there is a correction factor that is applied to determine the start and end of the events that is a result of the optimal threshold being closer to the peak of the bound population in the velocity distribution that is a result of the bound population having a narrower peak (smaller SD) since myosin stiffness is greater than the trap stiffness. After event identification, events were ensembled averaged as by synchronizing events in the x and y dimensions by applying linear regressions to the baseline prior to the start of the event (when the bead is moving at constant velocity) and overlaying events at the point where the linear regression intersected the start of the event as ID’d prior in analysis. The resulting ensembles could be fit with a model consisting of a linear portion that described the delay before the powerstroke and a double exponential reflecting the kinetics/mechanics of myosin going through the first and second powerstroke. Ensemble averages are typically calculated separately for the three populations of event durations that are prevelant in UFFC experiments (short, intermediate, and long) since these events represent different mechanochemical schemes of an actomyosin interaction.
3.4 Step-by-step single molecule event identification
This is a “step-by-step” walk through of the Hidden-Markov/Changepoint Analysis we use to analyze our single molecule laser trap data and includes everything on the journey from raw data to analyzed trace and everything on the way…buckle up. This section is a lot cooler if you are reading online and includes the R code to reproduce this by hand. You can download the data here
3.4.1 Raw data
Here is a raw data trace. This is unprocessed data as-is from the trap computer. The data is relative position of the bead in mV over time:
Figure 3.1: Raw trap data.
The data record is 89.1576 seconds long and has an average position of 6.2968527 mV.
3.4.2 Processed Data
The first step of the analysis is removing the “baseline” by centering the mean around 0. This can either be done by simply subtracting the baseline mean from every data point or by performing a piecewise linear detrend on the whole record. The latter accomplishes two things: 1) Centers mean around 0 and 2) removes any drift (i.e. wander correction). Additionally, in the lasertrapr app you can find the average baseline position by using a mean variance transformation of the data to select the baseline population or by selecting a quiescent period of the data where no binding events occur to calculate mean baseline position. Here we will detrend the data and convert from mV to nm using a “Step Calibration.” The step calibration is performed by moving the microscope stage a known distance, say 200 nanometers, and measuring the resulting change in the mV signal. We then can estimate the number of nanometers per mV.
Figure 3.2: Example of a step calibration. Stage was moved 200nm
The mV-to-nm conversion calibrated around the time this data trace collecting we are analyzing know was 30 nm/mV. We can convert our raw data from mV to nm and detrend the data (or visa versa).